Many organizations have employed the concepts of ‘asset specificity’ and ‘uncertainty’ from transaction cost theory (TCT), and ‘strategic resources’ from the resource-based view (RBV) as drivers of their information systems (ISs) sourcing decisions. They, however, face a dilemma when TCT and the RBV suggest different sourcing alternatives. The study identifies contexts where sourcing decisions made based on these two theories differ, and examines which theory accounts better for an organization's sourcing decisions. Results show that a high-specificity asset is a major driver of sourcing decisions. It overpowers the effects of uncertainty on sourcing decisions; while a non-strategic resource has no impact on sourcing decisions. In particular, where the two theories make conflicting predictions, organizations should not always outsource non-strategic resources. Non-strategic resources that involve high specificity should be retained internally. The paper concludes with implications for academics and practitioners, and a discussion of future research directions.
During the last decade, the e-commerce revolution has forced the transformation of traditional IS outsourcing structures into new IS outsourcing service configurations, e.g. Internet service outsourcing, application service outsourcing, and business process outsourcing. The IS outsourcing market worldwide is expected to increase to more than US$110 billion by 2006 (IDC, 2002). Outsourcing, however, does not always lead to competitive advantages and cost savings (Gerwig, 1999, Lacity and Hirschheim, 1993, Loh and Venkatraman, 1992 and Mahoney and Pandian, 1992), and the chances of success in outsourcing are at best 50/50 (Gartner Group, 2002). A wrong sourcing decision results in lost capabilities and exposure to risks that can result in business failure (Loh and Venkatraman, 1992 and Ngwenyama and Bryson, 1999). The ultimate drivers of sourcing decisions are not identical in all circumstances (Lacity and Hirschheim, 1993 and Nam et al., 1996) further complicating the sourcing decision. Organizations select the drivers of sourcing decisions based on the particular set of circumstances they face.
Two theories that have been used to understand and explain ITS sourcing decisions are transaction cost theory (TCT) (e.g. Ang and Straub, 1998; Aubert et al., 1996a and Aubert et al., 1996b; Lacity and Hirschheim, 1995 and Lacity et al., 1995) and the resource-based view (RBV) (e.g. Insinga and Werle, 2000, Prahalad and Hamel, 1990, Roy and Aubert, 2001 and Teng et al., 1995). TCT posits that organizations insource when the costs of using the market are higher than internal governance costs (Ngwenyama and Bryson, 1999 and Williamson, 1979). It builds upon three principal attributes of transactions: asset specificity, frequency, and uncertainty (Williamson, 1985b). The RBV, on the other hand, posits that organizations insource when a resource is strategic so as to enable them to sustain competitive advantage (Barney, 1991 and Mahoney and Pandian, 1992). It builds upon four properties of a strategic resource: value, rareness, imperfect imitability, and non-substitutability (Barney, 1991). A question that follows, which has yet to be empirically investigated, is ‘Are there contingencies that suggest when organizations should use either TCT or the RBV for their sourcing decisions?’
This study addresses the fundamental question: Which of TCT or the RBV better accounts for an organization's sourcing decision, where the organization deems that it has made an effective decision? The study develops a framework to identify circumstances where TCT and the RBV make different predictions about an organization's sourcing decision. A case-study approach is used to test these competing predictions in the context of an email-marketing system, which is an example of a contemporary business process outsourcing (BPO) decision. As a prediction from the two theories about an organization's sourcing decision may result in a mistaken decision, this study focuses on organizations that made an effective sourcing decision.
The study results offer guidance to practice, suggesting circumstances in which either TCT or the RBV should be employed for organizations' sourcing decisions. This study also fills a gap in the literature because no previous work has used empirical research to compare how well TCT and the RBV predict sourcing decisions.
The data from the three cases provide evidence on how the theoretical constructs
influence sourcing decisions and which theory, TCT or the RBV, better accounts for the
company’s sourcing decisions. The findings indicate that TCT explains an
organization’s sourcing decisions better than the RBV when these two theories make
conflicting predictions about a sourcing decision. In particular, a high-specificity asset
has a major impact on sourcing decisions. It overpowers the effects of uncertainty on
organizations’ sourcing decisions while a non-strategic resource has no impact on
sourcing decisions.
The study has not tested all the propositions. Low-specificity, strategic resources did
not exist in the study’s domain. Another study with domains involving this type of
resources is needed to test the predictions of sourcing decisions under the two theories.
The interview data (example 2, Section 5.1) implied that an organization’s ability to
predict the future outcome (manifestation of environmental uncertainty) might associate
with frequent changes (manifestation of transaction frequency). Future research might
examine how the interaction effect of environmental uncertainty and transaction
frequency affect an organization’s sourcing decision. This study emphasized the effects
of asset specificity on an organization’s sourcing decisions, but it was found that
specific and specialized assets are different. Future research studies might examine how
specialized assets affect an organization’s sourcing decisions. In addition, this study
focused on small, eastern culture firms as opposed to large, western culture firms for
which TCT and the RBV have been developed. It may limit transferability of the
study’s results. A number of future research studies might examine these two theories
in the same context.
